Assessing the Health Benefits of a Paris-Aligned Coal Phase out for South Korea

ASSESSING THE HEALTH BENEFITS OF A PARIS-ALIGNED COAL PHASE OUT FOR SOUTH KOREA

May, 2021

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

AUTHORS

Gaurav Ganti Anne Zimmer

Andreas Anhaüser* Charlotte Plinke

Lauri Myllyvirta** Carley Reynolds

Deborah Ramalope Matthew Gidden

Bill Hare

We would like to express our gratitude to the team of Solutions for Our Climate for their review and guidance in shaping this document.

A digital copy of this briefing along with supporting appendices is available at:

www.climateanalytics.org/publications

CITATION AND ACKNOWLEDGMENTS This publication may be reproduced in whole or in part and in any form for educational or non-profit services without special permission from Climate Analytics, provided acknowledgment and/or proper referencing of the source is made.

This publication may not be resold or used for any commercial purpose without prior written permission from Climate Analytics. We regret any errors or omissions that may have been unwittingly made.

This document may be cited as: Climate Analytics (2021). Assessing the Health Benefits of a Paris-Aligned Coal Phaseout for South Korea

* Affiliation: Greenpeace ** Affiliation: Center for Research on Energy and Clean Air

Supporting science based policy to prevent dangerous climate change enabling sustainable development

www.climateanalytics.org

In collaboration with:

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2

Key findings

● To contribute to the achievement of the Paris Agreement, South Korea needs to phase out coal from its electricity sector before 2030. The country’s 9th Basic Plan for Electricity Power Supply and Demand (9th BPESD) presents a unit-level operation schedule for coal power plants that would see nearly 27 GW of coal-fired power capacity still online in 2034, with coal eventually being phased out in 2054, almost 25 years later than is required to be Paris Agreement compatible.

● In this work we present two unit-level decommissioning schedules that are aligned with a Paris Agreement compatible CO2 emission reduction pathway. Both of these schedules require 4.2 GW of coal capacity to be retired each year, and units currently under construction would only be able to operate for four years at the most.

● Coal-fired power plants are a significant source of air pollution in South Korea, which has been linked to premature deaths due to increased risk for cardiovascular diseases, chronic and acute respiratory diseases, and other health impacts such as pre-term births and depression.

● Following the two Paris-compatible decommissioning schedules presented in this study could halve the number of premature deaths linked to air pollution from South Korean coal plants within the next 5 years and save over 18,000 lives (over 12,000 lives within South Korea) until the end of their operation, when compared to the current policy plan of phasing out coal in 2054.

Avoided health impacts under accelerated coal phase out (Market Scenario) Total Domestic Abroad

Outcome 95%-confidence interval 95%-confidence interval 95%-confidence interval best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 18,482 12,034 25,589 12,619 8,168 17,363 5,863 3,866 8,226 deaths Years of potential life 339,294 219,361 474,593 228,161 147,813 314,663 111,133 71,548 159,930 lost Preterm births 1,734 839 1,842 830 402 882 904 437 960 Asthma: New 3,261 706 7,373 2,552 552 5,787 709 154 1,586 cases Table: Avoided impacts (selected) in a Paris Agreement consistent coal phase out scenario compared to current policies.

● A corresponding reduction of over 1,700 preterm births (800 within South Korean boundaries), 3,000 new asthma cases (2,500 within South Korean boundaries) and more than 4.6 million work absence days (2.8 million within South Korean boundaries) is estimated to result from the accelerated decommissioning schedule as compared to current policy plans (see Table above).

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 3

Introduction

South Korea relies heavily on coal for electricity generation (coal accounted for 29% of installed capacity in 2020)—a significant contributor to air pollution and its associated health impacts. South Korea consistently ranks high among the countries where air pollution levels recommended by the World Health Organisation are exceeded by a wide margin [1].

This has led the South Korean government to declare air pollution a ‘social disaster’ that necessitates emergency mitigation measures [2]. Despite this, the government plans to continue to build coal-fired power plants with nearly 7.3 GW of coal capacity still at various stages of construction and planning [3].

th Figure 1 Comparison between potential CO2 emissions from coal generation under the 9 BPESD and a Paris Agreement consistent emission reduction pathway. *see explanation in footnote 1

The continued expansion of coal capacity stands in stark contrast to the benchmark coal phase out date of 2029 (Figure 1) that is consistent with the achievement of the Paris Agreement temperature limit1 [4]. The 9th BPESD (Box 1) released in December 2020 lays out a decommissioning trajectory that would see over 30 GW of coal capacity still online in 2030.

Such a trajectory would lock South Korea into a future that is not only inconsistent with greenhouse gas emission reduction targets (with emissions from coal-fired generation peaking only around 2025), but also exposes its population to elevated immediate health risks due to air pollution.

In this brief, we present two alternate decommissioning schedules that are consistent with meeting the 2029 benchmark for a complete coal phase out and highlight the health impacts avoided due to accelerated decommissioning of coal-fired power units in South Korea.

1 Climate Analytics modelled a Paris Agreement consistent emission reduction pathway for coal generation in South Korea based on a downscaling model applied to the IEA’s Energy Technology Perspectives “Beyond 2°C Scenario“, taking into account historical emissions until 2019. This pathway has been assessed by Climate Analytics to have characteristics that are consistent with the Long-Term Temperature Goal of the Paris Agreement. The Paris Agreement establishes a global commitment to limit warming “well below 2°C” and to pursue efforts to limit warming to 1.5°C.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 4

Box 1 The 9th Basic Plan for Electricity Supply and Demand

The 9th Basic Plan for Electricity Supply and Demand

The Basic Plan for Electricity Supply and Demand is a biennial report that is prepared by the Ministry of Trade, Industry and Energy in accordance with Article 25 of the Electricity Business Act and Article 15 of the Enforcement Decree. The document presents the long-term supply and demand plan (for a period of 15 years) for electricity in South Korea.

The 9th plan (December 2020) aims to decrease coal and nuclear capacity while increasing natural gas and renewable energy capacities to meet growing demand. The following are some of the key characteristics of the 9th plan:

● 7.2 GW of coal capacity is set to come online in the next five years; ● Installed coal capacity is set to peak at 40.6 GW in 2024 and then decline to 29 GW by 2034; ● 24 units (12.7 GW) of coal-fired power plants to be converted to run on natural gas by 2034; ● Increase natural gas capacity to 58 GW and renewable energy capacity to 78 GW (59% solar and 32% wind) in 2034.

Unit-level decommissioning schedules

Based on the emission reduction pathway presented in Figure 1, we propose two different decommissioning schedules for South Korea’s coal generation units based on a method we have developed to construct a “priority order” for unit-level retirements [5]. The “priority order” is based on two different perspectives:

• Regulator perspective: In this perspective, the most carbon emission-intensive coal power units are prioritised for decommissioning. Put differently, the decommissioning date for each unit is determined primarily by the amount of CO2 emitted per unit of electricity generated and secondarily (to break ties between units that have the same emission intensity) by the highest Long Run Marginal Cost (LRMC).

● Market perspective: The primary objective of this approach is to reduce the overall national cost of the shutdown by prioritising the decommissioning of units with the highest generation costs2 per unit of electricity generated. We use estimates for the generation costs from the Carbon Tracker Initiative [6].

The decommissioning calculations are then performed step by step. For each year in which the sum of emissions from coal plants is above levels consistent with the long-term goal of the Paris Agreement, plants are decommissioned until the emissions are at or below this level.3 Further details about the calculation methods for the two perspectives are presented in Annex I.

2 We select the Long Run Marginal Cost (LRMC) in line with the recommendation from Carbon Tracker Initiative reports given that the South Korean electricity market is heavily regulated by the government. 3 These scenarios do not account for significant factors including the importance of each unit to grid stability and expected return on investment. They are, by design, stylised scenarios that are meant to provide an insight into how different strategies could differ (or lead to similar conclusions, as we highlight in this brief).

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 5

The average annual capacity reduction of coal-fired power is 4.2 GW in both perspectives, with the last units shutting down in 2029. The phase out date differs by less than 2 years between the two perspectives for 50 out of 67 units (i.e., 75% of all units). This indicates that, while the perspective matters to some extent in the precise retirement year of the units, the narrow window for the retirement of all units is the most important aspect to keep in mind. We provide information per unit and a description of some insights at the unit-level in Annex II.

Air pollution and health impacts of continued reliance on coal

Coal combustion is not only an issue because of CO2 emissions, but also because it generates large amounts of air pollutants which are linked to a number of damaging environmental and public health impacts. Since both stem from the same process – fossil fuel combustion – air pollution can be avoided by policy measures primarily aimed at reducing carbon emissions [7]-[11].

Fine particulate matter (PM2.5) is known to increase the risk of severe health conditions, such as cardiovascular disease, chronic and acute respiratory disease, lung cancer and premature births [12]- [15], among others. Sulphur dioxide (SO2) causes acid rain [16], toxic heavy metals such as mercury can damage the nervous system, including the brain [17] and nitrogen oxides (NOx) affect the respiratory system, while also contributing to the formation of additional particulate matter and harmful ozone (O3) [18].

While air pollution has been labelled a ‘social disaster’ by the South Korean government, the rhetoric has not been accompanied by sufficient policy measures [2]. In the recent past, the concentration levels of PM2.5 in Seoul has been around two times higher than the recommended limits in the guidelines of the World Health Organisation [19].

While there has been a substantial decrease in coal-related fine particulate matter emissions compared to 2018,4 the country’s 60 coal-fired power plants still emitted 3,527 tonnes of PM2.5- related air pollutants in December 2020 [20]. According to the Ministry of Environment, the progress in reducing air pollution emissions was achieved by shutting down old coal-fired power plants early, halting operation during seasons with high air pollution levels and enforcing emission standards for coal power plants [20], demonstrating that an accelerated coal phase out can be an effective approach to cutting air pollutants.

Air quality and health benefits of accelerated coal decommissioning schedules

The population exposed to air pollution from coal-fired power plants would benefit substantially from an accelerated, Paris Agreement compatible decommissioning schedule. In order to assess this quantitatively, we conducted a modelling study using a numerical weather model with a chemistry module to study the dispersion of pollutants emitted by the power plants. We present a brief overview of the method in Box 2 and a detailed overview in Annex III.

4 The Ministry of Environment reports a 60% decrease from the 8,781 tons emitted in December 2018 compared to December 2020 [20]. The actual effectiveness of the more recently adopted measures to manage air pollution from coal power plants, such as the most recent Comprehensive Plan on Fine Dust Management (CPFDM) implemented in November 2019 [25], remains to be seen when the economy recovers from COVID-19.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 6

Both schedules consistent with the Paris Agreement would result in a steep reduction of close to 79% in the estimated number of premature deaths resulting from air pollution from South Korean coal power plants when compared to the projection under the 9th BPESD (Figure 2a and b for total effect and 2e and f for domestic effect). While there is little difference between the two phase out scenarios, the potential reduction in accumulated effects between the current policy scenario and the accelerated decommissioning schedules is substantial. The number of premature deaths per year is almost halved in the next five years under these schedules, as opposed to an increase under the 9th BPESD. Cumulatively, around 18,000 premature deaths5 - more than 12,500 of these within South Korea - can be avoided in the Paris Agreement schedules as compared to the 9th BPESD (which sees more than 23,000 premature deaths until 2054, almost 15,900 of these within South Korea – see Figure 2c and d and Table 1-3).

Total a b

c d

4 4

5 Best estimate. 95%-confidence ranges are shown in Tables 1-3.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 7

Domestic e ff

g h

4 4

Figure 2: Estimated total premature deaths per year (figures a and b for total and e and f for domestic) and cumulative premature deaths accumulated over the time of the phase out (figures c and d for total and g and h for domestic) related to air pollution from South Korean coal power plants comparing the current policy trajectory of the 9th BPESD with the accelerated coal phase out scenarios in line with the Paris Agreement (Left – Market Perspective, Right – Regulator perspective). Shaded areas show 95%-confidence intervals.6

There is significant variation in the regional impacts of air pollution. We construct a hypothetical case overlaying the emissions of all operating and planned coal power plants to illustrate the regional per capita impacts7 (shown in Figure 3 as air pollution deaths per million inhabitants). Coal power related air pollution deaths are especially high in the vicinity of the power plants, which are grouped in three clusters: one in the South, one in the North-East and one in the West. The 10 most affected

6 The rate of premature deaths from coal-fired power plants (panel a, b, e and f) is changing over time due to two opposing effects competing against each other. 1) Sharp downward cuts appear when a power plant unit goes offline (often multiple units at a time). In the early years, some new units go online, which leads to a sharp increase. 2) These sharp sudden drops are superposed to an underlying constant increase in the death rate which is due to the changing demography: The population is projected to age, which makes people more vulnerable to air pollution-related deaths (see Annex II A2). The numbers in this figure are rounded – for underlying numbers please see tables I-III in Annex IV. 7 In reality, some power plant units are planned to already be taken offline before new ones currently in the pipeline go online. Thus, there is no time when all units are online concurrently. Yet, the map shows very clearly that there are regions that suffer the highest health impacts from coal power generation in relative terms (i.e. accounting for differences in population density).

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 8

municipalities (in terms of deaths per population if all units were online at the same time) are: Gwangyang, Boryeong, Suncheon, Donghae, Hongseong, Hadong, Gurye, Gokseong, Sacheon and Samcheok. In terms of absolute numbers, the most populous areas suffer the most premature deaths, as a higher population density also increases the number of people exposed to the air pollution from a particular coal power plant (not shown).

Figure 3: Regional distribution of coal power related air pollution deaths per capita when modelling air pollutant emission of all current and future coal power plants in South Korea combined (best estimate). Note that this shows a hypothetical case with all power plant units emitting at the same time to illustrate which regions are affected the most. In reality, some units would already be phased out when newly constructed power plants would come online.

On average, each of these premature deaths cut short a person’s life, leading close to 430,000 years of potential life lost due to air pollution from South Korea’s coal power fleet until their end of operation under current policies (Annex IV Table I). Over 280,000 of these potential life years are lost to residents of South Korea, the remaining more than 140,000 to people abroad. Under either of the Paris compatible phase out scenarios, these numbers are drastically decreased to around 90,000 years of potential life lost, saving more than 330,000 life years in total (Tables 1 and 2 as well as Tables I, II and III in Annex IV).

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 9

Table 1: Accumulated avoided health impacts of South Korean coal-fired power plants from 2021-2054, total, domestic and abroad, under the Regulator Perspective Scenario compared to the Current Policies Scenario.

Avoided health impacts under Regulator Scenario Total Domestic Abroad

Outcome 95%-confidence interval 95%-confidence interval 95%-confidence interval best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 18,410 11,993 25,480 12,527 8,113 17,232 5,883 3,880 8,248 deaths Years of potential 337,632 218,417 472,071 226,502 146,807 312,305 111,130 71,610 159,766 life lost Preterm births 1,728 836 1,835 827 400 879 901 436 956 Asthma: New 3,239 701 7,323 2,528 547 5,732 711 154 1,591 cases Asthma: Emergency room 7,307 4,544 10,047 4,122 2,581 5,650 3,185 1,963 4,397 visits Work absences 4,647,484 3,953,635 5,336,693 2,809,835 2,390,349 3,226,511 1,837,649 1,563,286 2,110,182 (person days)

Table 2: Accumulated avoided health impacts of South Korean coal-fired power plants from 2021-2054, total, domestic and abroad, under the Market Perspective Scenario compared to the Current Policies Scenario.

Avoided health impacts under Market Scenario Total Domestic Abroad

Outcome 95%-confidence interval 95%-confidence interval 95%-confidence interval best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 18,482 12,034 25,589 12,619 8,168 17,363 5,863 3,866 8,226 deaths Years of potential 339,294 219,361 474,593 228,161 147,813 314,663 111,133 71,548 159,930 life lost Preterm births 1,734 839 1,842 830 402 882 904 437 960 Asthma: New 3,261 706 7,373 2,552 552 5,787 709 154 1,586 cases Asthma: Emergency room 7,326 4,555 10,072 4,136 2,589 5,669 3,190 1,966 4,403 visits Work absences 4,652,626 3,958,009 5,342,597 2,819,235 2,398,346 3,237,305 1,833,391 1,559,663 2,105,292 (person days)

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 10

Air pollution also causes a variety of non-lethal negative health impacts. Under current policies, more than 2,300 preterm births (of these 1,100 in South Korea) between 2021-2054 are estimated to be attributed to air pollution from South Korean coal power, around three quarters of which would be avoided by either of the two phaseout scenarios (Annex IV Table 1). Under both phaseout scenarios, the number of future new asthma cases can be reduced by over 3,000 and the number of emergency room visits due to an exacerbated pre-existing asthma condition can be reduced by well over 7000 cases (Table 1 and Table 2).

All of these health impacts not only cause personal harm to the affected individual, but also entail economic losses to the society due to the need to use health care services, but also through loss of productivity, manifesting in more than 6.3 million work absences (person days) under current policies (see Table I in Annex IV). This number would be reduced by 73% to around 1.7 million under either of the phase out scenarios.

Box 2: Air pollution impact calculation methodology

Methodology

The unit-level decommissioning schedules serve as an input to estimate the respective air quality benefits of an accelerated South Korean coal phase out. Using the concentration-response functions for various health outcomes, we computed the health impacts of the continued operation of the power plants in the future. This captures the contribution of each coal unit to air pollutant concentration, which is then used to calculate impacts on human health for the exposed population within South Korea as well as neighbouring countries,8 taking projected population developments and changes in age structures into account. We assess these impacts in comparison to the expected health impacts under the 9th BPESD. Further details on the methodology are presented in Annex III.

Conclusion

Our results show that an accelerated phase out of coal power generation in South Korea is not only needed to comply with the internationally agreed goals of the Paris Agreement, which South Korea has ratified, it would also save about 18,000 lives (over 12,000 domestically) compared to the current policy plan of phasing out coal in 2054.

In just five years’ time, implementing an accelerated coal phase out could halve the number of premature deaths per year stemming from air pollution from South Korean coal plants. The choice of the criteria for ranking which coal power plant units should be phased out in which order matters only slightly, with phasing out more carbon-intensive coal power plants first resulting in a slightly higher number of lives that could be saved compared to taking the marginal cost perspective for ranking coal units for phase out.

Our findings are corroborated by findings from recent scientific literature. Maamoun et al (2020) rank over 2,000 operating coal-fired power plants in the world based on their age, carbon intensity and air pollution potential, and identify the top plants to be retired early according to these criteria to be

8 China, Japan and North Korea.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 11

located in China, India and South Korea [22], supporting our findings for a need for an accelerated coal phase out in South Korea.

Quantifying the health benefits from reduced air pollution for South Korea achieving ambitious climate mitigation targets, Kim et al (2020) estimate that the economic health benefits could considerably outweigh the total costs of climate change mitigation in South Korea [23]. They estimate savings from reduced health expenditures to be about 0.14 billion USD and those from reducing the number of work hours lost to 0.38 billion USD, while the valuation of avoiding premature deaths alone could reduce economic health costs by about 23 billion USD applying a value of statistical life approach. A recent study focussed on coal in the power sector found that the cumulative cost of health impacts under the 9th BPESD could be as high as 21 billion USD [26].

It is important to note that the results presented in this brief only consider the air pollution and health impacts from the coal-fired power plants themselves. In this regard, the presented health benefits of an accelerated coal phase out can be considered conservative for two reasons:

● First, the processes along the supply chain for coal, including mining, hauling, storage and coal ash disposal also cause considerable air pollution and health impacts that are not included in our estimates. ● Moreover, the 9th BPESD foresees that 24 coal power plants are transformed into natural gas power plants. Recent research has demonstrated that a coal-to-gas transition in the electricity system is unlikely to be consistent with achieving the warming limit of the Paris Agreement [27], and instead direct transition to renewables may be a better pathway. Natural gas also contributes to air pollution impacts, which are not considered here.

● As a consequence, the total air pollution as well as CO2 emissions stemming from power generation under the business-as-usual scenario can be expected to be even higher, and thus the health and climate benefits of an accelerated coal phase out (without transforming coal power into natural gas power plants) would be even higher.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 12

References

[1] MOTIE, “보도/해명 | 산업통상자원부 홈페이지,” 2020. [Online]. Available: http://www.motie.go.kr/motie/ne/presse/press2/bbs/bbsView.do?bbs_cd_n=81&bbs_seq_n=163670. [Accessed: 07-Mar-2021]. [2] The Guardian, “‘Social disaster’: South Korea brings in emergency laws to tackle dust pollution,” 13-Mar- 2019. [3] C. Shearer et al., “Boom and Bust 2020: TRACKING THE GLOBAL COAL PLANT PIPELINE,” 2020. [4] Climate Analytics, “South Korea must exit coal by 2029 to be in line with the Paris Agreement,” Climate Analytics, 20-Feb-2020. [Online]. Available: https://climateanalytics.org/latest/south-korea-must-exit- coal-by-2029-to-be-in-line-with-the-paris-agreement/. [Accessed: 04-Aug-2020]. [5] Climate Analytics, “A Stress Test for Coal in Europe under the Paris Agreement,” 2017. [6] M. Gray and D. D’Souza, “BROWN IS THE NEW GREEN: Will South Korea’s commitment to coal power undermine its low carbon strategy?,” Carbon Tracker Initiative Analysis Report, 2019. [Online]. Available: https://carbontracker.org/reports/south-korea-coal-power/. [Accessed: 07-Mar-2021]. [7] C. Oberschelp, S. Pfister, C. E. Raptis, and S. Hellweg, “Global emission hotspots of coal power generation,” Nat. Sustain., vol. 2, no. 2, pp. 113–121, Feb. 2019, doi: 10.1038/s41893-019-0221-6. [8] E. Amster, “Public health impact of coal-fired power plants: a critical systematic review of the epidemiological literature,” Int. J. Environ. Health Res., pp. 1–23, Oct. 2019, doi: 10.1080/09603123.2019.1674256. [9] K. Mahavong, P. Pataranawat, and S. Chinwetkitvanich, “MERCURY CONTAMINATION IN ENVIRONMENT SURROUNDING COAL-FIRED POWER PLANT,” doi: 10.21660/2017.33.2581. [10] N. L. Mills et al., “Adverse cardiovascular effects of air pollution,” Nature Clinical Practice Cardiovascular Medicine, vol. 6, no. 1. Nature Publishing Group, pp. 36–44, 25-Nov-2009, doi: 10.1038/ncpcardio1399. [11] S. Zhou et al., “Impact of a Coal-Fired Power Plant Shutdown Campaign on Heavy Metal Emissions in China,” Environ. Sci. Technol., vol. 53, no. 23, pp. 14063–14069, Dec. 2019, doi: 10.1021/acs.est.9b04683. [12] A. Vodonos, Y. A. Awad, and J. Schwartz, “The concentration-response between long-term PM 2.5 exposure and mortality; A meta-regression approach,” Environ. Res., vol. 166, pp. 677–689, Oct. 2018, doi: 10.1016/j.envres.2018.06.021. [13] I. Braithwaite, S. Zhang, J. B. Kirkbride, D. P. J. Osborn, and J. F. Hayes, “Air pollution (Particulate matter) exposure and associations with depression, anxiety, bipolar, psychosis and suicide risk: A systematic review and meta-analysis,” Environmental Health Perspectives, vol. 127, no. 12. Public Health Services, US Dept of Health and Human Services, 18-Dec-2019, doi: 10.1289/EHP4595. [14] P. Klepac, I. Locatelli, S. Korošec, N. Künzli, and A. Kukec, “Ambient air pollution and pregnancy outcomes: A comprehensive review and identification of environmental public health challenges,” Environmental Research, vol. 167. Academic Press Inc., pp. 144–159, 01-Nov-2018, doi: 10.1016/j.envres.2018.07.008. [15] F. Liu et al., “Associations between long-term exposure to ambient air pollution and risk of type 2 diabetes mellitus: A systematic review and meta-analysis,” Environmental Pollution, vol. 252, no. Pt B. Elsevier Ltd, pp. 1235–1245, 01-Sep-2019, doi: 10.1016/j.envpol.2019.06.033. [16] S. Kumar, “Acid Rain-The Major Cause of Pollution: Its Causes, Effects,” 2017. [17] S. Zhou et al., “Impact of a Coal-Fired Power Plant Shutdown Campaign on Heavy Metal Emissions in China,” Environ. Sci. Technol., vol. 53, no. 23, pp. 14063–14069, Dec. 2019, doi: 10.1021/acs.est.9b04683. [18] United States Environmental Protection Agency, “Basic Information about NO2 | Nitrogen Dioxide (NO2) Pollution | US EPA.” [Online]. Available: https://www.epa.gov/no2-pollution/basic-information-about- no2#Effects. [Accessed: 15-Mar-2021]. [19] D. Trnka, “POLICIES, REGULATORY FRAMEWORK AND ENFORCEMENT FOR AIR QUALITY MANAGEMENT: THE CASE OF KOREA-ENVIRONMENT WORKING PAPER No. 158.” [20] Ministry of Environment (MOTIE), “Joint update of the environmental authorities of Korea and China on their fine dust response.,” Press Release, 2021. [Online]. Available: http://eng.me.go.kr/eng/web/board/read.do?pagerOffset=0&maxPageItems=10&maxIndexPages=10& searchKey=&searchValue=&menuId=461&orgCd=&boardId=1432680&boardMasterId=522&boardCate goryId=&decorator=. [Accessed: 26-Feb-2021].

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

[21] Solutions for Our Climate, “기후솔루션,” 2020. [Online]. Available: http://www.forourclimate.org/sub/data/view.html?idx=12&curpage=2. [Accessed: 15-Mar-2021]. [22] N. Maamoun, R. Kennedy, X. Jin, and J. Urpelainen, “Identifying coal-fired power plants for early retirement,” Renew. Sustain. Energy Rev., vol. 126, 2020, doi: 10.1016/j.rser.2020.109833. [23] S. E. Kim et al., “Air quality co-benefits from climate mitigation for human health in South Korea,” Environ. Int., vol. 136, no. June 2019, p. 105507, 2020, doi: 10.1016/j.envint.2020.105507. [24] IPCC, “Stationary Combustion,” in 2006 IPCC Guidelines for National Greenhouse Gas Inventories, H. S. Eggleston, L. Buendia, K. Miwa, T. Ngara, and K. Tanabe, Eds. 2006. [25] Ministry of Environment (MOTIE), “Air pollution.” [Online]. Available: http://eng.me.go.kr/eng/web/index.do?menuId=464. [Accessed: 09-Mar-2021]. [26] CREA. (2021). The Health and Economic Cost of Coal Dependence in South Korea’s Power Mix. [27]. Ganti, G., Brecha, R., Strefler, J., Uekcerdt, F., Luderer, G., & Gidden, M. (2021). Coal-to-Gas Bridge Incompatible with Paris Agreement Goals. doi: https://doi.org/10.21203/rs.3.rs-289099/v1

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2

Annex I – Methodology for unit-level phase out schedules and emission calculations

In this study we consider two approaches to determine the phase out schedule at the unit level:

● “Regulator” perspective: it adopts an environmental integrity approach and prioritises the shutdown of the least efficient units, while also taking into account the maximisation of the revenue they can generate. For this perspective, we assume generation units are sorted primarily according to their carbon intensity (amount of CO2 emitted per unit of electricity generated). To reduce the overall economic loss of the phase out and given that many generation units have similar carbon intensity characteristics, a second sorting is applied where priority for phase out is given to the units with lower LRMC in each of the carbon intensity ranges. We assume a constant capacity factor of 70% over time for our scenarios (further details on the emission calculation below). ● “Market” perspective: it aims to reduce the overall national cost (regardless of region) of the shut down for investors and owners by keeping units with higher economic value online as long as possible. Similar to the Regulator perspective, the sorting of the units is done using a two-step approach. Units are sorted according to the LRMC and then ties are broken using the emission intensity of electricity generation. The shutdown is performed in a step-by-step manner. For each year in which the sum of emissions from the coal plants are above target emissions pathway, plants need to be shut down until the emissions are below this level. Coal power units are sorted as explained above and those units with highest priority will be shut down in a certain year, as depicted in Figure 13.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

Figure 13 – Schematic overview of methodology. Each of the boxes labelled A to G shows emissions from a power unit. The blue line indicates cost optimal coal emissions pathways in line with the Paris Agreement Long-Term Temperature Goal, and t0 through t4 depict the time steps (years). If we assume that our shutdown regime starts in t1, this means that plants G and F need to shut down – as indicated by the grey colour. In t2 plant E needs to be shut down under a least cost strategy and in t3 only plants A and B may remain in operation. In t4 all remaining plants need to be shut down, as emissions need to reach zero.

To estimate emissions resulting from currently operating and planned coal power plants in South Korea we used the Global Coal Plant Tracker (GCPT) database, which provides information on every known coal-fired power generation unit, including its location, status, investor, capacity, combustion technology9 and fuel, year of opening and planned retirement. We merge the data with data from KEPCO and assign a capacity factor of 70% to all units. The data used in this report comprise of detailed information per plant concerning the country, its capacity, status and combustion technology, which allows to estimate CO2 from these plants, using the following formula:

Yearly emissions: 1 ��!" = ��! ∗ ∗ ��!" ∗ ��! ∗ � ��! with:

�� are the yearly emissions of plant unit i in Mt CO2 in a particular year.

�� is the Capacity of plant unit i in MWel. MWel describes the electrical output of a power plant (unit). About two-thirds (actual value depending on the combustion technology) of the energy contained in a coal power plant’s fuel is lost while converting it into electricity. The thermal energy released during the conversion is usually not used anymore but gotten rid of via cooling towers or rivers.

�� is the conversion efficiency of a power plant unit: How much of the energy contained in the fuel (coal) is converted to electricity. In general, this is higher for modern plants, which dominate South Korea’s coal power generation.

�� is the load factor of the power plant in a particular year. The load factor is the ratio of the actual power plant output over its theoretical maximum output and is usually calculated over the course of a year. The theoretical maximum output can be calculated by assuming that a power plant runs at its nameplate capacity 24 hours a day, 365 days a year i.e. a power plant unit with a capacity of 100 MW has a theoretical maximum output of:

100 MW ∗ l0 ∗ ayr = 876.000 MWh. Actual output over a given year is lower since the plant will always operate at full output – e.g. due to demand fluctuations – and has to be taken offline completely for maintenance. There is uncertainty around future utilisation rates of coal power plants in South Korea, and hence we select a default value of 70% for our calculations. It is important to note that this would not impact the emission intensity of electricity generation that are used in the regulator scenario.

�� is the emissions factor, which contains information on how much CO2 is released for a given amount of coal burned. Unit is kg CO2/TJ. Higher-grade coal contains a higher share of carbon, which is converted to CO2 during combustion. We use emission factors from (IPCC, 2006). Since this source

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2

contains only emission factors for pure types of coal, we assumed a 50/50 share for plants that use two different coal grades, e.g., bituminous and sub-bituminous coal.

� is a conversion factor to end up with the correct units (Mt CO2/yr).

9 The database distinguishes between different combustion technologies in the following categories: subcritical, supercritical and ultra-supercritical without or with CCS, ranking from least to most efficient respectively. We do not consider coal-fired power plants retrofitted with CCS technology further in our analysis.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 3

Annex II – Unit-level phase out schedules

Emission Policy Regulator Market Intensity LRMC Unit Name Retirement Retirement Retirement (g CO2 / (Unit) kWh)

Boryeong #1 2020 2020 2020 1045.552 50.66 Boryeong #2 2020 2020 2020 1045.552 50.54 Donghae #1 2028 2020 2020 1256.639 55.36 Donghae #2 2029 2021 2020 1166.901 54.32 Honam #1 2021 2021 2021 1175.02 49.16 Honam #2 2021 2021 2021 1175.02 49.16

Samchonpo #1 2021 2021 2021 1012.987 54.15

Samchonpo #2 2021 2021 2021 1012.987 53.9 Taean #2 2025 2022 2021 939.6995 52.92

Samchonpo #4 2024 2021 2021 972.3855 52.3

Samchonpo #3 2024 2021 2021 972.3855 51.4 Dangjin #3 2030 2023 2021 861.3834 51.1 Yeosu #1 2046 2022 2021 881.4979 50.76 Taean #4 2029 2022 2021 939.6995 50.45 Dangjin #4 2030 2023 2021 861.3834 49.97

Gangreung Anin #1 2052 2026 2022 795.0736 51.63

Gangreung Anin #2 2052 2026 2022 795.0736 51.63 Taean #3 2028 2022 2022 939.6995 49.6 Dangjin #1 2029 2024 2022 861.3834 49.53

Gosung Hai #1 2051 2026 2022 795.0736 49.43

Gosung Hai #2 2051 2026 2022 795.0736 49.43

Samcheok Greenpower #2 2047 2027 2022 795.0736 49.35 Yeosu #2 2041 2021 2023 1094.004 49.27 Taean #1 2025 2022 2023 939.6995 49.15 Boryeong #3 2043 2021 2023 969.909 49.01

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

Dangjin #9 2045 2028 2023 770.3096 48.92

Yeongheung #1 2034 2024 2023 861.3834 48.65 Dangjin #2 2029 2024 2023 861.3834 48.59 Dangjin #8 2037 2024 2024 847.3219 48.51 Dangjin #7 2037 2025 2024 847.3219 48.48

Yeongheung #2 2034 2024 2024 861.3834 48.25 Dangjin #10 2046 2029 2024 770.3096 48.19 Taean #10 2047 2029 2024 770.3096 48.14

Bukpyeong #2 2047 2023 2024 867.3625 48.1

Bukpyeong #1 2047 2023 2024 867.3625 48.06 Boryeong #4 2043 2021 2024 969.909 47.93 Taean #7 2037 2025 2024 847.3219 47.82 Boryeong #5 2025 2021 2025 969.909 47.53 Boryeong #6 2025 2022 2025 969.909 46.82 Dangjin #6 2036 2023 2025 874.5617 47.73 Taean #5 2032 2024 2025 861.3834 47.73 Hadong #1 2026 2022 2025 939.6995 47.72 Hadong #2 2027 2022 2025 939.6995 47.72 Dangjin #5 2035 2023 2025 874.5617 47.72 Hadong #3 2028 2022 2026 939.6995 47.66

Yeongheung #5 2044 2028 2026 783.0674 47.64

Yeongheung #6 2044 2028 2026 783.0674 47.62 Hadong #4 2028 2024 2026 861.3834 47.6 Hadong #5 2031 2024 2026 861.3834 47.53 Hadong #6 2031 2024 2026 861.3834 47.47

Samchonpo #5 2027 2022 2027 969.909 44.34

Shin-Seocheon 2051 2027 2027 795.0736 47.51 Boryeong #8 2038 2023 2027 874.5617 47.49

Samcheok Greenpower #1 2046 2027 2027 795.0736 47.41

Sinboryeong #2 2047 2027 2027 795.0736 47.25

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2

Samchonpo #6 2028 2022 2028 969.909 43.62

Samcheok Blue Power #1 2054 2025 2028 808.2415 47.18

Samcheok Blue Power #2 2054 2025 2028 808.2415 47.18

Yeongheung #4 2038 2029 2028 770.3096 47.15

Sinboryeong #1 2047 2028 2028 795.0736 47.08 Taean #8 2037 2025 2029 847.3219 46.97 Taean #9 2046 2029 2029 770.3096 46.85 Taean #6 2032 2024 2029 861.3834 46.79 Boryeong #7 2038 2023 2029 874.5617 46.55

Yeongheung #3 2038 2029 2029 770.3096 46.54 Hadong #7 2038 2025 2029 847.3219 45.88 Hadong #8 2039 2029 2029 770.3096 45.83

On average, the Market scenario presents a phase out schedule that takes place 11.5 years sooner than that of current policy, whereas the Regulator scenario sees it come forward by 11.9 years. There is not a significant difference between the two. Here, we present some key insights:

a. We first explore the units that retire earlier in the Regulator scenario as compared to the Market scenario: i. Samcheonpo 6 and Boryeong 7 retire 6 years earlier. ii. Samcheonpo 5 and Taeahn 6 retire 5 years earlier. iii. Hadong 3 and 7, Boryeong 5 and 8 and Taeahn 8 retire 4 years earlier. iv. Hadong 1 and 2, Samcheok Blue Power 1 and 2, Boryeong 4 and 6 retire 3 years earlier. v. Apart from Samcheok Blue Power 1 and 2 (that are equipped with ultrasupercritical boilers), all units are equipped with supercritical boilders. b. We now explore the units that retire earlier in the Market scenario as compared to the Regulator scenario: i. Taeahn 10, Samcheok Green 2, Dangjin 9 and 10 retire 5 years earlier. ii. Gosung Hi 1 and 2, Gangreung Anin 1 and 2 retire 4 years earlier. iii. These units are equipped with ultrasupercritical boilers. Gangreun Anin 1 and 2 are under construction, while all other units are either operating or near completion. c. In the market scenario, the units under construction (including Gosung Hi 1 and 2, Gangreung Anin 1 and 2, Samcheok Blue Power 1 and 2, and Shinseocheon) would

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 3

retire almost immediately after their completion – in the regulator scenario they would operate for at most four years. This demonstrates the high risk that these units face of becoming stranded assets, even if completed.

i. Among the ultrasupercritical units completed around 2016-2017, such as Bukpyeong 1 and 2, Dangjin 9 and 10, Taeahn 10, Shinboryeong 1 and 2, Samcheok Green 1 and 2, those located along the East coast (Samcheok Green 2, Bukpyeong 1 and 2, etc.) would need to retire relatively sooner. ii. Among the existing supercritical /subcritical units, Yeosu 1 and 2, Boryeong 3 and 4, Yeonheung 5 and 6, Taeahn 7, Dangjin 7 and 8 were identified as the ones that would need to be retired sooner than the current policy schedule. Several factors could be considered in the prioritisation of these, such as recent equipment investment made in those units, etc. d. In the Regulator scenario, the retirement sequence is almost identical to the order of construction years. This sheds light on a potential path to an accelerated coal phase out process. Aging units that are still kept due to various policy concerns, for example Donghae 1 and 2, could be pushed forward. Then, units located in the vicinity of others with an expedited retirement schedule, such as Samcheonpo 3 and 4, Boreyong 3,4,5, and6, Taeahn 1,2,3 and 4, and Hadong 1,2 and 3, could also be retired earlier. i. Among the new units, Bukpyeong 1 and 2 have considerably higher pollutant emissions in our calculations, and this can be validated and confirmed by an investigation of the causes.

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 4

Annex III – Methodology air pollution and health impacts

A.1 Atmospheric dispersion modelling system The atmospheric dispersion model consists of two major components. A meteorology module is used to simulate the regional meteorological conditions around the power plants (A.1.1). This is combined with a chemistry-transport model to study the propagation of the power plant emissions to the environment (A.1.2). A.1.1 Meteorology module

Figure A.1: A numerical weather model is run on three nested domains (red boxes) centred around South Korea. • Model. The meteorology around the power plants is modelled using version 3 of the The Air Pollution Model (TAPM). Although TAPM includes the ability to model pollutant dispersion, only the meteorology component of TAPM is used. • Domains. The meteorology module is run on three nested 75x75 grids centred in South Korea with spatial resolutions of 20 km, 10 km and 5 km, respectively, getting finer towards the centre (Figure A.1). • Boundary conditions are derived from the GASP model data of the Australian Bureau of Meteorology. • Spin up and analysis period. We ran the meteorology module for one year (model time). Since we analysed the multi-year health impact, a recent yet representative year was selected. In each simulation, the model is run for the last seven days of 2015 and the whole year of 2016. The first seven days are used to let the model spin up, and only the 2016 data are used for the analysis. • Selection of the model year. In order to choose a year which is as representative for the long run as possible, we performed a rough statistical analysis of meteorological surface observations between 2005-2018. We analysed wind speed, wind direction, temperature, humidity, rainfall, cloud base height and sky cover from standard airport weather observations

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

through NOAA’s ISD data and screened for a year in which the values are close to average and none are the extreme values for the past 10 years. Based on this criterion, we selected 2016 as representative model year. A.1.2 Chemistry-transport module • Model. The atmospheric dispersion, chemical transformation and deposition of the power plant emissions is modelled by version 7 of the CALPUFF model. • Receptor array. The chemistry-transport module is run on an irregular array of 11,224 receptors centred around each of the power plants. The spatial resolution of the receptors ranges from 700m in the immediate surrounding of the power plant and 20km at the edge of the outermost domain of the meteorology module. • Other pollution sources. As we are solely focusing on the impacts from the power plants, no emission sources other than the studied power plants are included in the model. The influence of other sources of emissions on the chemical transformation of power plant emissions is taken into account indirectly by inputting the monthly concentrations of background chemical species (NH3, O3, H2O2) from Koplitz et al (2017) into CALPUFF. • Pollutant species. Modelled are the power plants emissions of mercury (elemental, divalent and particle-bound), NO, NO2, SO2 and primary PM2.5. • Background concentrations of O3, NH3 and H2O2 are included for use by the chemistry module. • Output. The model outputs a time series of near-surface concentrations of the pollutants for analysis to gridded receptor locations across the model domains. • Accounting for operation time. The model is run for the whole year at the full-operation emissions rates. The resulting ground-level pollutant concentration fields are used as such for assessing maximum short-term air quality impact. For the purposes of health impact assessment (Section A.2), the average concentrations are scaled down by the plant’s projected load factor, effectively spreading the plant’s annual emissions volume evenly through the year. Power plant geometry and emission data sources

The pollutant emission rates and flue gas release characteristics used for the modelling are based, as far as possible, on self-reported data by power plant owners and government sources. Other publicly available information was used in the analysis. Over 90% of data was collected from the Korean government’s information disclosure system and data reported by the power plant owners to the parliament member’s office of the National Assembly. The following parameters have been used to determine power plant emission characteristics; • Reported annual emissions from the power plant owners and government statistics • Emission limit values or average stack emission concentrations • Amount of actual or expected air pollutant (PM2.5, NO2, SO2) emissions • Unit age, capacity and thermal efficiency • Annual operating hours • Calorific value of coal • Flue gas volume flow (FGV) • Flue gas exit speed and temperature • Stack height • Stack inner diameter The location of the power plants was sourced from visual inspection of satellite imagery provided by the Google map platform by and from environmental impact assessment documents published by the power plant owners. The stack inner diameter was measured from high-resolution satellite imagery, when actual information was not available from regulatory sources or literature. When information on stack height and flue gas temperature was not available, median values in the South Korean dataset for units with the same size were used. Flue gas release velocity was calculated from stack diameter

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2

and flue gas flow, when not available. When annual emissions volumes were not reported, they were calculated as: SFGV * CAP / EFF * CF * FGC, where SFGV is specific flue gas volume (Nm3/GJ thermal input), estimated as the median for plants for which flue gas flow rate, capacity and thermal efficiency were known; CAP is electric capacity, EFF is thermal efficiency, CF is the annual average capacity factor and FGC is pollutant concentration in flue 3 gas, given in ppm for SO2 and NOx and mg/Nm for dust. As mercury emissions are not reported by plant operators, they were calculated using average mercury emission rate per tonne of coal burned in South Korea calculated from AMAP/UNEP Global Mercury Assessment. The power plant and emission data shown in Table A1 is used as the basis of modelling the plants’ air quality impacts using the CALMET-CALPUFF modelling system.

Table A1. Stack and flue gas characteristics and baseline emission rates of the modelled power plants (metric units). Stack and flue gas parameters are reported by the operator. Emission data is (a) reported by the operator or (b) computed from expected capacity and assumptions on thermal efficiency and capacity factor, as explained above. Stack Flue gas Emissions

exit Unit lat lon height diameter temperature SO2 NOx PM Plant speed data source (deg) (deg) (m) (m) (°C) (m/s) (t/a) (t/a) (t/a)

Yeosu 1 34.8404 127.6917 150 4.8 91 27.2 128 392 14 a

Yeosu 2 34.8404 127.6917 150 4.8 91 21.6 76 427 11 a

Yeongheung 1 37.2429 126.4463 200 6.6 90 18.1 1239 937 60 a

Yeongheung 2 37.2429 126.4463 200 6.6 90 18.1 1363 975 52 a

Yeongheung 3 37.2429 126.4463 198 6.3 90 20.4 888 536 16 a

Yeongheung 4 37.2429 126.4463 198 6.3 90 20.4 819 513 25 a

Yeongheung 5 37.2429 126.4463 200 6.9 95.3 17.4 497 450 22 a

Yeongheung 6 37.2429 126.4463 200 6.9 95.3 17.4 507 463 20 a

Taean 1 36.0004 126.2451 150 8.83 79.83 8.9 320 560 53 a

Taean 2 36.0002 126.2443 150 8.83 77.87 8.9 272 341 31 a

Taean 3 35.9999 126.2435 150 8.83 78.41 8.9 322 835 94 a

Taean 4 35.9997 126.2427 150 8.83 77.69 8.9 691 710 99 a

Taean 5 35.9991 126.2413 150 5.4 78.49 23.4 993 886 78 a

Taean 6 35.9991 126.2411 150 5.4 82.22 23.4 1353 1037 93 a

Taean 7 35.9986 126.2396 150 5.4 95.87 27.7 451 567 20 a

Taean 8 35.9986 126.2395 150 5.4 96.41 27.7 598 700 24 a

Taean 9 35.9972 126.2373 150 7.3 91 32.9 706 795 77 a

Taean 10 35.9973 126.2375 150 7.3 91 32.9 1011 816 53 a

Shin Boryeong 1 36.3840 126.4850 150 7.5 90 16.8 1150 507 29 a

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 3

Shin Boryeong 2 36.3840 126.4850 150 7.5 90 16.8 1005 763 45 a

Samcheonpo 1 34.9117 128.1092 200 5.3 119 37.5 687 1122 39 a

Samcheonpo 2 34.9117 128.1092 200 5.3 109 37.5 557 1178 57 a

Samcheonpo 3 34.9117 128.1092 200 5.3 95 37.5 1230 2035 75 a

Samcheonpo 4 34.9117 128.1092 200 5.3 100 37.5 1044 1302 71 a

Samcheonpo 5 34.9117 128.1092 200 5.3 139 24.2 645 521 118 b

Samcheonpo 6 34.9117 128.1092 200 5.3 148 23.8 726 586 133 b

Samcheok Green Power 1 37.1860 129.3400 90 8.8 90 15.1 595 1445 105 a

Samcheok Green Power 2 37.1840 129.3400 90 8.8 90 15.1 330 1008 60 a

Honam 1 34.5114 127.4404 150 7 90 6.6 681 862 17 a

Honam 2 34.5114 127.4406 150 7 90 6.6 1379 1558 32 a

Hadong 1 34.9500 127.8200 150 9.3 83 7.4 1113 1094 50 a

Hadong 2 34.9500 127.8200 150 9.3 83 6.6 1174 706 47 a

Hadong 3 34.9510 127.8200 150 9.3 83 6.6 1128 707 49 a

Hadong 4 34.9520 127.8190 150 9.3 83 6.5 970 855 30 a

Hadong 5 34.9520 127.8190 150 9.3 83 6.7 898 721 40 a

Hadong 6 34.9530 127.8190 150 9.3 83 6.8 819 1129 40 a

Hadong 7 34.9540 127.8180 150 5.4 82 18.6 720 808 37 a

Hadong 8 34.9540 127.8180 150 5.4 82 19.1 886 855 49 a

Donghae 1 37.2907 129.0847 150 4 154 16.2 770 262 7 a

Donghae 2 37.2909 129.0845 150 4 154 16.2 857 341 10 a

Dangjin 1 37.0315 126.3051 151 6.5 85 15.4 691 602 52 a

Dangjin 2 37.0315 126.3048 151 6.5 85 15.4 662 706 53 a

Dangjin 3 37.0316 126.3045 151 6.5 85 15.4 562 592 44 a

Dangjin 4 37.0317 126.3042 151 6.5 85 15.4 664 544 42 a

Dangjin 5 37.0318 126.3037 150 5.4 90 22.2 540 687 53 a

Dangjin 6 37.0319 126.3034 150 5.4 90 22.2 627 667 54 a

Dangjin 7 37.0320 126.3031 150 5.4 90 22.2 393 464 42 a

Dangjin 8 37.0320 126.3028 150 5.4 90 22.2 397 431 47 a

Dangjin 9 37.0323 126.3023 200 7.4 91 25.7 796 829 22 a

Dangjin 10 37.0324 126.3018 200 7.4 91 25.7 1079 1058 27 a

Bukpyeong 1 37.4770 129.1459 150 5.3 90 24.5 1212 1910 84 a

Bukpyeong 2 37.4770 129.1459 150 5.3 90 24.5 1212 1910 84 a

Boryeong 1 36.4020 126.4880 150 8.864 85 7.2 570 959 53 a

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 4

Boryeong 2 36.4020 126.4890 150 8.864 85 7.2 470 1320 40 a

Boryeong 3 36.4020 126.4900 150 8.788 90 8.8 516 416 94 b

Boryeong 4 36.4020 126.4910 150 8.762 90 7.3 804 464 26 a

Boryeong 5 36.4020 126.4920 150 8.788 90 7.3 1013 667 49 a

Boryeong 6 36.4020 126.4930 150 8.788 90 7.3 1146 689 34 a

Boryeong 7 36.4020 126.4940 150 5.4 90 17.2 235 312 43 a

Boryeong 8 36.4020 126.4940 150 5.4 90 17.2 229 604 32 a

Shin Seocheon 1 35.2394 126.5016 150 7.4 90 21.0 1364 1159 238 b

Goseong Hi 1 34.9020 128.1230 190 7.5 90 22.5 1500 1275 261 b

Goseong Hi 2 34.9020 128.1230 190 7.5 90 22.5 1500 1275 261 b

Gangneung Anin 1 37.7340 128.9780 102 7.6 101 21.9 1389 1121 254 b

Gangneung Anin 2 37.7340 128.9780 102 7.6 101 21.9 1389 1121 254 b

Samcheok Blue power 1 37.4070 129.1770 250 7.4 90 22.3 1415 1203 247 b

Samcheok Blue power 2 37.4070 129.1770 250 7.4 90 22.3 1415 1203 247 b

A.2 Health impact assessment The results of the near-surface concentration of air pollutants emitted by the power plants as modelled by the atmospheric dispersion modelling system are used to assess the exposure of the human population to air pollution and its impact on public health.

A.2.1 Health outcomes The number of people affected by negative health outcomes caused by the excess pollution have been assessed using empirical values of relative risks relating negative health outcomes (asthma symptoms, low birth weight, death) to increases in pollutant concentrations. The relative risk r expresses how much more likely an individual is to experience that health outcome if they (or their mother) are exposed to a certain excess pollution compared to if they were not exposed:

mx / m0 = r, (1) where mx is the incidence rate under the increased pollution Δx, and m0 is the incidence rate in absence of the excess pollution. The incidence rate is the number of cases where an individual experiences this health outcome per number of the relevant population group (asthmatic children, newborns, total population). In some epidemiological models, r(x) is given as an empirical function. In others, r(x) is an

, exponential function for mx << 1: r = exp(c Δx), (2) with c being a constant called concentration response factor. Combining Eqs. (1) and (2) gives

mx = m0 exp(c Δx). Since the number of cases is the number of the relevant population group P times the incidence rate, the number of cases under the higher pollutant concentration is

dx = P m0 exp(c Δx).

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 5

The number of cases attributable to the excess pollution is

Δd = dx - d0= P m0 [exp(c Δx) - 1]. Integrating Δd spatially over the model domain gives the total number of cases attributable to the excess pollution within the model domain.

Data sources for the health impact assessment • Population (present and future). Data on total population and population age structure were taken from the GBD project 2019 (IHME 2020). • Background incidence rates (present and future). o Baseline death rates and years of life lost for South Korea were taken from the GBD project 2019 (IHME 2020). o Background incidence rates for other health impacts were taken from the same sources as in Myllyvirta (2020). o The health impacts are adjusted by age group-specific changes in population and all- cause mortality, based on historical data and projections in UNPD World Population Prospects 2019 (medium variant). • Background pollution. The baseline concentrations of PM2.5 and NO2 were taken from van Donkelaar et al. (2016) and Larkin et al. (2017), respectively. • Administrative domain borders are taken as defined in version 3.6 (May 2018) of the GADM project. These boundaries and those shown on any maps solely reflect the data source used and do not imply recognition or support to any party where there may be territorial disputes. • Concentration response functions (CRFs). CRFs have been used from the same sources as in Myllyvirta (2020), except that NO2-related mortality is taken from Faustini et al (2014) and SO2- related mortality from Krewski et al (2009) (see Table A2).

Table A2. Concentration response functions.

Concentration- Age response Concentration No-risk group Effect Pollutant function change threshold Reference Incidence data

1-18 New asthma NO2 1.26 (1.10 - 10 ppb 2 ppb Achakulwisut Achakulwisut et cases 1.37) et al. 2019 al. 2019

3 3 0-17 Asthma PM2.5 1.025 (1.013– 10 μg/m 6 μg/m Zheng et al. Anenberg et al. emergency 1.037) 2015 2018 room visits

3 3 18-99 Asthma PM2.5 1.023 (1.015– 10 μg/m 6 μg/m Zheng et al. Anenberg et al. emergency 1.031) 2015 2018 room visits

3 3 Newborn Preterm birth PM2.5 1.15 (1.07, 10 μg/m 8.8 μg/m Trasande et Chawanpaiboon 1.16) al. 2016 et al. 2019

3 25-99 Deaths from PM2.5 non-linear 2.4 μg/m Burnett et al. IHME 2020 non- 2018 communicable diseases and lower respiratory infections

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 6

3 0-4 Deaths from PM2.5 non-linear 5.8 μg/m IHME 2020 IHME 2020 lower respiratory infections

3 25-99 Deaths from PM2.5 non-linear 5.8 μg/m IHME 2020 IHME 2020 lower respiratory infections

3 3 25-99 Premature NO2 1.04 (1.02– 10 μg/m 4.5 μg/m Faustini et al. IHME 2020 deaths 1.06) 2014

25-99 Premature SO2 1.02 (1.014– 5 ppb 0.02 ppb Krewski et al. IHME 2020 deaths 1.026) 2009

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 7

Annex IV – Health impacts across the three scenarios

Here, we present the results on the estimated health impacts in more detail, showing the estimates for the different scenarios.

Table I: Accumulated health impacts of South Korean coal-fired power plants from 2021-2054, total, domestic and abroad, under current policies (9th Basic Plan for Electricity Power Supply and Demand).

Current policies

Total Domestic Abroad

Outcome 95%-confidence interval 95%-confidence interval 95%-confidence interval best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 23,332 15,188 32,309 15,880 10,273 21,857 7,452 4,915 10,452 deaths Years of potential life 428,223 276,760 599,087 287,127 185,890 396,110 141,096 90,870 202,977 lost Preterm births 2,348 1,136 2,493 1,128 546 1,198 1,220 590 1,295 Asthma: New 4,382 948 9,909 3,451 746 7,825 931 202 2,084 cases Asthma: Emergency 9,793 6,083 13,471 5,572 3,484 7,642 4,221 2,599 5,829 room visits Work absences 6,376,941 5,424,894 7,322,620 3,905,690 3,322,604 4,484,870 2,471,251 2,102,290 2,837,750 (person days)

Table II: Accumulated health impacts of South Korean coal-fired power plants from 2021-2054, total, domestic and abroad, under the Regulator Perspective Scenario.

Regulator Scenario

Total Domestic Abroad

95%-confidence interval 95%-confidence interval 95%-confidence interval Outcome best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 4,922 3,195 6,829 3,353 2,160 4,625 1,569 1,035 2,204 deaths Years of potential life 90,591 58,343 127,016 60,625 39,083 83,805 29,966 19,260 43,211 lost Preterm births 620 300 658 301 146 319 319 154 339 Asthma: New 1,143 247 2,586 923 199 2,093 220 48 493 cases

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 1

Asthma: Emergency 2,486 1,539 3,424 1,450 903 1,992 1,036 636 1,432 room visits Work absences 1,729,457 1,471,259 1,985,927 1,095,855 932,255 1,258,359 633,602 539,004 727,568 (person days)

Table III: Accumulated health impacts of South Korean coal-fired power plants from 2021-2054, total, domestic and abroad, under the Market Perspective Scenario.

Market Scenario

Total Domestic Abroad

Outcome 95%-confidence interval 95%-confidence interval 95%-confidence interval best low high best low high best low high estimate estimate estimate estimate estimate estimate estimate estimate estimate Premature 4,850 3,154 6,720 3,261 2,105 4,494 1,589 1,049 2,226 deaths Years of potential life 88,929 57,399 124,494 58,966 38,077 81,447 29,963 19,322 43,047 lost Preterm births 614 297 651 298 144 316 316 153 335 Asthma: New 1,121 242 2,536 899 194 2,038 222 48 498 cases Asthma: Emergency 2,467 1,528 3,399 1,436 895 1,973 1,031 633 1,426 room visits Work absences 1,724,315 1,466,885 1,980,023 1,086,455 924,258 1,247,565 637,860 542,627 732,458 (person days)

Assessing the health benefits of a Paris-aligned coal phase out for South Korea 2